Slicer for getting data

[sundayDriver]

Hi,

I wanted to first say that power BI is an awesome tool that everyone in the team highly enjoys!

I wanted to ask the following question:

Our team has been running into the following problem:

We have the same data (that is too much for power BI to load) and we want to use the same report for them.

Same way as you can create a slicer that contains which data to be displayed is there an easy way that you can configure where the data will be pulled from?

To give an example you might have a large chunk of data of type blue, green and yellow.

Is there a way that we can on the fly select blue, green or yellow and get the appropriate data as we can't load data for all the categories?

[Phil_Seamark · Accepted solution]

Hi @sundayDriver,

The limit for PBIX files is currently 1GB.  Power BI uses compression that can compress up to around 10x so is your source data really around 10GB?  

There are some tricks to reduce the import file size such as

• remove unnessary columns,
• aggregate your data in your source data if possible
• Split Datetime data into 2 columns (if time is needed)

If these don't help then sure, DirectQuery is an option but hopefully your source system is SQL2016 with columnstore indexes optimised to the max 🙂

[v-sihou-msft · Accepted solution]

@sundayDriver

As @Phil_Seamark and @Greg_Deckler, for large amount of data, you'd better use DirectQuery mode due to the 1 GB limitation. If you need to filter expected records (blue, green or yellow) during data retrieval, you need to do it in your query.

Regards,

[Phil_Seamark · Accepted solution]

Hi @sundayDriver,

The limit for PBIX files is currently 1GB.  Power BI uses compression that can compress up to around 10x so is your source data really around 10GB?  

There are some tricks to reduce the import file size such as

• remove unnessary columns,
• aggregate your data in your source data if possible
• Split Datetime data into 2 columns (if time is needed)

If these don't help then sure, DirectQuery is an option but hopefully your source system is SQL2016 with columnstore indexes optimised to the max 🙂

[Greg_Deckler]

You would want to look at using DirectQuery.